This work investigates the hypothesis that a possible cause of poor health amongst the socially disconnected is increased stress levels. The relatively new theory of allostatic load (AL) posits that the body's response to stress can come at a cost, which accumulates over time and eventually leads to negative health outcomes. There is increasing evidence that high levels of AL are predictive of poor health states, but few studies have attempted to demonstrate that elevated AL results from a stressful life history.
This dissertation analyzes data from the Social Environment and Biomarkers of Aging Study (SEBAS), a nationally representative survey conducted in Taiwan of those aged 54 and older in 2000. Using neuroendocrine biomarkers and social variables, a combination rarely found in large surveys of this quality, this work tests whether various measures of social disconnection and social stress lead to riskier biomarker levels in an AL construct. Key independent variables I use in the analysis include widowhood status and its duration, living alone, not living with a married son, and subjective responses to a number of questions probing the respondents' stress history. The dependent variable used here is called neuroendocrine allostatic load (NAL), which is an index combining four neuroendocrine biomarkers (cortisol, DHEAS, epinephrine, norepinephrine). Among other reasons for using this construct is that the included markers are core to the physiological stress response and AL theory, and have hitherto been little studied.
A major finding of this work is that, contrary to expectation, major and enduring stressors appear to have no influence on NAL levels. In contrast, psychosocial reports of current stress (amongst women) are associated with measurable increases in NAL, even though AL theory would predict negligible impact (since any additional AL stemming from recent events would be dwarfed by events having occurred over the entire life course). Thus, although baseline levels of the neuroendocrine markers may still be useful in predicting worse health outcomes, this dissertation calls into question whether the levels truly represent accumulated costs of past insults. Indeed, the baseline neuroendocrine markers may well be capturing current, ephemeral states at the time of the interview.
My Research